Simulation tools for model-based robotics: Comparison of Bullet, Havok, MuJoCo, ODE and PhysX

By Tom Erez, Yuval Tassa, E. Todorov. 2015

358 citations recorded by the source corpus.

Subfields: Computational Theory and Mathematics, Control and Systems Engineering, Mechanical Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Control and Systems Engineering

In Mechanical Engineering

Citations in the corpus

Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations (2017, background): “MuJoCo: A physics engine for model-based control.”
DeepMind Control Suite (2018, background): “It compares favourably to other popular engines (Erez et al., 2015), especially for articulated, low-to-medium degree-of-freedom (DoF) models in contact with other bodies.”
Learning to Solve a Rubik’s Cube with a Dexterous Hand (2019, background): “ics [24] and is suitable for rich-contact hand manipulation tasks.”
Brax - A Differentiable Physics Engine for Large Scale Rigid Body Simulation (2021, methodology): “Non-Brax data was adapted with permission from the authors of [42] and plotted here for comparison.”
Model Identification via Physics Engines for Improved Policy Search (2017, background): “There are many examples of popular physics engines that are becoming increasingly efﬁcient [1]– [6].”
SceneNet: An annotated model generator for indoor scene understanding (2016, background): “Moreover, SceneNet could also be integrated with a physics engine [21] and any robotics simulation engine [22] to allow further understanding of scenes beyond geometry and semantics.”
Simulation-based design of dynamic controllers for humanoid balancing (2016, background): “Modern physical simulators [25] solve the governing dynamics equation with LCP constraints to accurately model contacts.”
MuJoCo HAPTIX: A virtual reality system for hand manipulation (2015, background): “We have recently shown [2] that MuJoCo outperforms a number of alternative simulators in terms of both speed and accuracy on modelling systems relevant to robotics, especially simulated hands grasping objects.”
Fast Model Identification via Physics Engines for Data-Efficient Policy Search (2017, methodology): “This paper presents an approach for model identiﬁcation by exploiting the availability of off-the-shelf physics engines [Erez et al. , 2015] for simulating robot and object dynamics in a given scene.”
dm_control: Software and Tasks for Continuous Control (2020, methodology): “It is written in Python , exploiting the agile workﬂow of a dynamic language, while relying on the C -based MuJoCo physics library, a fast and accurate simulator (Erez et al., 2015), itself designed to facilitate research (Todorov et al., 2012).”
A novel DDPG method with prioritized experience replay (2017, methodology): “MuJoCo [21], [22] is used as the simulator and OpenAI Gym is adopted to load the MuJoCo model.”